Electrical, Thermal, and Morphological Properties of Poly (ethylene terephthalate)-Graphite Nanoplatlet Nanocomposites

Graphite nanoplatelets (GNP) were incorporated with poly(ethylene terephthalate) (PET) matrix by melt-compounding technique using minilab compounder to produce PET-GNP nanocomposites, and then the extruded nanocomposites were compressed using compression molding to obtain films of 1 mm thickness. Percolation threshold value was determined using percolation theory. The electrical conductivity, morphology, and thermal behaviors of these nanocomposites were investigated at different contents of GNP, that is, below, around, and above its percolation threshold value. The results demonstrated that the addition of GNP at loading >5 wt.% made electrically conductive nanocomposites. An excellent electrical conductivity of ~1 S/m was obtained at 15 wt.% of GNP loading. The nanocomposites showed a typical insulator-conductor transition with a percolation threshold value of 5.7 wt.% of GNP. In addition, increasing screw speed enhanced the conductivity of the nanocomposites above its threshold value by ~2.5 orders of magnitude; this behavior is attributed to improved dispersion of these nanoparticles into the PET matrix. Microscopies results exhibited no indication of aggregations at 2 wt.% of GNP; however, some rolling up at 6 wt.% of GNP contents was observed, indicating that a conductive network has been formed, whereas more agglomeration and rolling up could be seen as the GNP content is increased in the PET matrix. These agglomerations reduced their aspect ratio and then reduced their reinforcement efficiency. NP loading (>2 wt.%) increased degree of crystallinity and improved thermal stability of matrix slightly, suggesting that 2 wt.% of GNP is more than enough to nucleate the matrix.


Introduction
Graphite nanoplatlets (GNP) which is called also as expanded graphite (EG), or exfoliated graphite. In graphite form, the graphene sheets are held together by van der waals interaction which allows some molecules, atoms or ions to be intercalated between these sheets. Usually, GNP prepared by modified of graphite using graphite intercalation method (chemical exfoliation method). In this method, graphite is immersed in a concentrated mixture of sulfuric acid with other oxidizers agents. The resulting products are called as graphite intercalation compounds (GIC) or as expandable graphite. Typically, the GIC materials are heated to a high temperature of ~ 1000 o C for few seconds that generates huge amounts of gases such as sulfur dioxide and water resulting in extra expansion of graphene sheets; which is known as GNP, exfoliated graphite or expanded graphite (EG). This procedure is achieved under nitrogen or in air atmospheres. Further exfoliation of these materials can produce a single atomic layer of graphite which known as graphene. Typically, high ultrasonication process will be used for breaking down the GNP to smaller sheets. Both GIC and GNP are commonly modified graphite forms that have used for preparing polymercarbon composites [1][2][3][4].
During the previous years, polymer-GNP nanocomposites have been widely investigated to develop materials to be used where both electrical, thermal and mechanical properties are required. According to a review by Li at el., [2] and an investigation study by Kamran at el, [5], many researchers prepared various conducting polymer nanocomposites containing various GNP weight fractions and sizes under different preparing conditions. The most results indicate significant enhancement in both mechanical, thermal and electrical properties of final nanocomposites. Although, One of the main problems for polymer nanocomposite materials is the appropriate dispersion and distribution of Nano fillers into the matrix that needed to be addressed before taking the full advantages of particulate nano-scale fillers .
Furthermore, PET is a thermoplastic polymer. It used in textile fibers, films and packaging technologies. This because of that PET has good properties such as mechanical, chemical resistance, ability to be spun, thermal and dimensional stabilities [6,7].
Although the above-stated properties, improvement of its electrical conductivity is necessary for electronic applications [8]. For example, replacing of indium tin oxide (ITO) electrodes as these materials have poor mechanical and lower conductivity when compared to the conductive carbon nano-fillers such as graphene nanoplatlets or carbon nanotubes [2]. The electrical conductivity value higher than 10 -8 S/m are required for numerous electronic applications. For instance, semiconducting, antistatic and electromagnetic shielding materials for electronic devices [9].
To achieve electrically conductive pathways within the polymer composites with high conductivity value and acceptable mechanical properties and process-ability, a reduction of the amount of conductive fillers is required. Therefore, percolation theory is frequently used to determine the percolation threshold value which is the minimum content of conductive fillers associated with the conversion in the electrical behavior of the composite from insulate to conductive phase. This theory also describe the insulator--conductor transitions in polymer composites made of a conductive fillers and an insulating polymer matrix. Previous published studies have been shown that the percolation threshold value strongly depends on the aspect ratio (length-to-diameter) of the filler particles [10][11][12].
There are three common adopted methods for the production of composites: solution, in situ polymerization and melt compounding. The latter method is an environmentally friendly technique. Therefore, it was used in this study. Relatively, there are few investigations reported on PET-GNP nanocomposites. For instance, a study by Zhang et al, [4] prepared PET-GNP nanocomposites using melt blending. A homogenous dispersion of GNP in the PET matrix was noted in this study after some shearing has been used to improve interaction between PET matrix and GNP. Similar work done by Li et al, [8] who prepared PET nanocomposites with diverse concentrations of GNP using melt mixing method, reported good dispersed GNP within the PET matrix, leading to improvements in the mechanical, electrical and thermal properties of nanocomposites compared with the unfilled PET. The electrical percolation threshold was found to be ~ 5 wt.% Al-jabareen et al, [13] added GNP into PET by melt mixing to enhance its oxygen barrier properties. The authors observed a uniform distribution and dispersion of GNP into PET. The results indicated an increase in crystallinity, thermal stability and Young's modulus of the resultant composite compared to pure PET matrix. However, tensile strength and elongation at break showed reductions of 56 % and 40 %, respectively. A mother work carried out by Paszkiewicz et al, [14] on the electrical conductivity of PET-GNP nanocomposites prepared by suit polymerization approaches, proved that good dispersion of GNP in the PET matrix. The transition from insulator to conductor system was start at amount of 0.05 wt. % GNP.
Such low a percolation threshold related to the large surface area, high aspect ratio and uniform dispersion of the GNP as well as manufacturing method used which monomers polymerized in the existence of the GNP, therefore allowing stronger interactions.
Although the above-mentioned studies have examined properties of PET-GNP nanocomposites, only a few studies have characterized their properties after determine the minimum amount of GNP required for forming a conductive pathways within the PET-GNP nanocomposites system. Therefore, in this study, a nanocomposite was prepared using PET as polymer matrix and GNP as nanoscale filler. The effect of GNP on electrical conductivity, morphology and thermal behavior of PET matrix investigated as a function of GNP loadings. At first stage of this study, the percolation threshold (c) value was determine using percolation theory. This value is important for improving fabrication of the polymer composite (PCs) and their properties. The development of PCs based on conductive filler has concentrated on reducing the c in order to reduce cost, improve the fabrication or process-ability, and enhance mechanical properties of the final composite products [9]. Therefore, finding such value is very important to indicate the performance of the final composite products.

Preparation of nanocomposites
Both as-received materials of PET and GNP dried at about 120 ºC in vacuum oven overnight before melt compounding was carried out. The PET-GNP nanocomposites were prepared through the melt compounding technique by mixing or/and compounding PET and GNP in a molten state using a Thermo-Haake Minilab compounder. The melt compounding technique has described in details somewhere else [1]. In the current study, the time of mixing was 5 minutes at processing temperature of 280 ºC and speed of screws was 45 rpm. Then extruded samples were compressed at the same processing temperature i.e. 280 ºC for 10 min using compression molding machine and followed by quenching in ice bath to reduce the crystallinity and then chopped, dried at 40 ºC for 24 h and stored for the further characterization and testing.

Characterization of nanocomposites 2.3.1 Electrical conductivity and percolation threshold measurements
An impedance spectroscopy, model NumetriQ PSM1735, was used for determining both electrical conductivity values (σ) and percolation threshold values () for PET and its nanocomposites. Samples Plates of dimension (~10x10x1 mm 3 ) were cut from compressed films and then coated using sliver paint. The coating is to reduce the contact resistance between the samples and copper wires that attached to the specimen using silver epoxy adhesion. Resistance of all tested samples was measured in a range of 1 Hertz up to 10x10 -6 Hertz at 1 volte as an amplitude voltage. For checking the accuracy, each sample run five times at room temperature.

Morphology observations
The morphology state of GNP and the PET nanocomposites was characterized using a SEM (Model: Philips SEM XL30, at an accelerating voltage of 10 -20 kV). The nanocomposites were cryogenically frozen in a liquid nitrogen for 5 min and then fractured. The samples mounted on pin stubs using adhesive tape and then coated with a layer of gold using an Edwards S150B sputter coater, to avoid any charge during scan observation.
TEM (Philips CM200) was applied also at an accelerating voltage of 200 kV.
Nanocomposites samples were embedded into epoxy resin for slices purpose using an ultra-microtome and then diamond blade was used to cut slices of ~ 50 nm thick from the mid of the nanocomposites. addition, GNP were mixed with ethanol, sonicated using an ultrasonic bath at room temperature for 20 min and a little drops were dropped onto a copper grid using a micropipette. The grid was left to dry in a fume cupboard to evaporate the solvent.

Thermal behaviors
Crystallization behaviors of PET and its nanocomposites were examined using DSC (TA Instrument DSC Q-100). Samples were scanned from room temperature 270 o C under nitrogen condition using a heat-cool-heat run at a heating and cooling rate of 10 o C/min. To confirm results, different three specimen of each samples were measured.
Thermal stability was investigated using a thermogravimetric analyzer (TGA, TA Q-500). Samples were scanned from room temperature to 700 o C under nitrogen and air gas condition.

Electrical conductivity and percolation threshold measurements
From impedance spectroscopy measurements, the σ values are obtained as a function of frequencies. In this study, frequency at 10 Hz have chosen to obtain the values of σ for comparative purpose for all samples. This is because of that at low frequencies, insulator-conductor transitions can be only noticed [11].
Where, σ is the electrical conductivity (S/m), L is the distance between the electrodes (m), and A is the cross sectional area (m 2 ) and R is the measured electrical resistance (Ω). Figure 1 displays the values of σ of PET matrix and its nanocomposites as a function of GNP loadings. It is clear that the σ of the nanocomposites having < 5 wt. % of GNP is nearly constant and close to the σ of pure PET. Then, a great increase in σ occurs between 5 and 9 wt. % of GNP, this is followed by a small improvement in the σ with increasing the amount of GNP. Thus, the  of these nanocomposites is expected between 5 and 9 wt. % loadings of the GNP. Typically, the variation in the conductivity of PCs as a function of the conductive filler content exhibits an S-shaped curve for transition from insolating to conductive system [16]. Therefore, the PET-GNP nanocomposites exhibited a typical transition curve (figure 1) in this study.
To illustrate transitions of insulator to conductor phase or behavior of PCs reinforced with conductive fillers, percolation theory is applied. This theory will assist the determination of dimensions of conductive network in polymer matrices and percolation threshold values of PCs system. According to this theory, the σ of the PCs is calculating by following equation [11][12][13][14][15][16][17][18][19]: Where σ is coefficient constant, c is percolation threshold value (vol. %),  is the filler loading (vol. %) and t is dimension of the conductive network path. In theory, t values of 1.33 and 2.2 represent 2D and 3D systems, respectively. However; experimental values are also stated outside this range somewhere else [17,18]. Due to in fact that density of the nano-filler (such as GNP) is frequently an approximate value.
Therefore, weight fractions have used for determining the values of  and t in numerous published studies [12,15,16]. In this study, the all determinations will be in expression of weight fractions.
According to percolation theory, the best-fit data to the percolation threshold model (Equation (2)

Figure 1 Electrical conductivity of PET-GNP nanocomposites. The inset plot is log σ versus log (-c) for determination of percolation threshold value.
Generally, in PCs system, high aspect ratio of such nano-filler found to reduce the  values. However, based on the data sheets that provided by the supplier, the calculated aspect ratio for GNP ~ 1875. Therefore, value of  expected to be lower than the value obtained in these studies. The reasons for this could be due to that the GNP has unreduced graphite oxide in the commercial products and then some functional groups still on its surface after the preparation process. Therefore, FTIR technique was used in the current study to identify the functional groups attached to the surface of GNP during exfoliation techniques ( Figure 2). This result is in agreement with those one obtained by Li et al, [8] who reported some functional groups on the surface of the GNP used. However, a good an attraction can occur between the ester groups (-C=O) of PET (see Figure 3) and functional groups such as hydroxyls acids (-OH) on the surface of GNP which leads to improvements in the desperation state of these nanoparticles. In addition, there is a possibility of an agglomeration and rolling up of GNP filler during melt compounding process, this could be reduce aspect ratio of GNP sheets.
Additionally, t ~ 4.72 was much higher than the theoretical values mentioned earlier.

Similar value (t ~ 4.25) has been obtained by Zhang et al,[4] for PET-EG
nanocomposites. The authors proposed that the reasons for this deviation; i.e. tunneling transmission, the microstructure and anisotropic behavior of the composites.
Additionally, the screw speed and residence time during melt process are expected to impact the values of σ of PCs [20,21]. Therefore, in the present study, the impact of these two processing parameter on the σ was investigated. Two samples were selected just before and after the c value of the nanocomposites to note the effect of processing parameter on its conductivity.

Morphological characterization
The morphology of GNP was examined before its incorporation into the PET matrix in order to understand their size and shapes. Figure 4 shows SEM image of as received GNP. As can be seen, it shows the separated platelets resulting from break down of expandable graphite particles during the preparation process of GNP [22].

Figure 4 SEM image of GNP.
Similar morphologies for GNP were observed in previous studies [8,22,23].  c, showed some rolling up and agglomeration ( figure 6), indicating that a conductive network has been formed which allows electron to move.

Figure 6 SEM images at low and high maginfications of fracture surfaces of PET-GNP nanocomposites containing 6 wt.% of GNP.
More agglomeration could be observed in figure 7 as the GNP content is increased.
Similar morhplogy observations have been reported for polyproplene nanocomspoites by Kalaitzidou et al, [24]. In addition, Zhang et al, [4] who considered the relationship between the morphology and the σ of PET-EG nanocomposites. In the SEM results, it has been observed that the adding of EG above the c, particle contacted with each other, indicating that conductive path has been developed. However, below the c, the conductive particles were observed far away from each other, thus, there is no path for electron to move in the nanocomposite system.

Crystallization behavior
Pure PET and the nanocomposites with varying contents of GNP (2, 4, 6, 8 and 10 wt. %) were exposed to a heat/cool/heat cycles at heating and cooling of 10 o C/min. The DSC curves in figure 9(a) -(d) show cold crystallization peaks, melting peaks from the first run, melt crystallization peaks and melting peaks from the second cycle, respectively. The data obtained from these peaks are presented in table 2. and Tm-Tmc), indicating that 2 wt. % loading of GNP is more than enough to nucleate the PET matrix. Such an effect in the crystallizations of PET is in agreement with previous work that done by [25,26]. The Xc can be calculated using the enthalpies of both crystallization and melting, according the following equation [27,28]. The Xc of PET and its nanocomposite was calculated from the first-heat run as these data reflect the thermal history of the samples. Where ∆H is the melting enthalpy (J/g), ∆H is the cold crystallization enthalpy (J/g), ∆Ho is the theoretical enthalpy of 100 % crystalline PET (∆H =140 J/g) [26] and is the weight fraction of GNP.
The addition of GNP was found to affect the Xc as presented in table 2. There is an improvement of ~ 7.5 % in the value of Xc upon adding of 2 wt. % of GNP. However, additional adding of GNP does not show any considerable alteration in the values of Xc. Similar behavior was observed by Al-jabareen et al, [13], who studied PET-GNP nanocomposites, and reported an enhancement of 4 % in value of Xc at 1.5 wt. % GNP concentration. In contrast, it was reported that GNP did not change the values of Xc of a PP matrix [25].

Figure 9 DSC curves of PET-GNP nanocomposites showing that cold crystallization peaks (a), melting peaks from the first scan (b), subsequent cooling curves (c) and melting peaks from second heating run (d).
Furthermore, the addition of GNP appears to have no noticeable effect on values of Tm within the content of GNP that used in this study (table 2). Comparative peaks of melting for two the heating runs are shown in figure 9(b) and 9(d). From these curves and

Thermal Stability
Modified Graphite forms such as GNP have used to improve thermal stability of polymer matrices. For example, Al-jabareen et al, [13] who reported that the adding of GNP into PET matrix delay the onset temperature of degradation, indicating the improvement of thermal stability. The authors attributed this improvement to 2D structure of GNP that acts as barrier for oxygen diffusion through the nanocomposites.
A similar results was given by Li et al, [8]. In other studies, thermal stability enhancements have been reported for polycarbonate-GNP nanocomposites [29], using the same grade of GNP as used in this study; and when they used poly(methyl methacrylate) as a polymer matrix [30].

Figure 10 TGA thermograms of nanocomposites examined under nitrogen atmosphere
The results of the present study agree with the these studies, as in the current work, the incorporation of GNP into PET matrix enhanced its thermal stability of PET under nitrogen and air as shown in figures 10 and 11 (table 3).

Conclusion
In present study, a large number of nanocomposites were prepared using PET matrix and GNP nano-filler by melt-compounding method in order to determine the percolation threshold value for PET-GNP nanocomposites system. The electrical conductivity, morphology, thermal stability, crystallization behaviors and degree of crystallinity of these nanocomposites were characterized. In particular, at ~ 6 wt. % of GNP, an enhancement in conductivity was observed. Such loading marks the insulator conductor transition that having a percolation threshold value of 5.7 wt. %. The conductivity that is required for anti-static applications was observed just above the percolation threshold value. Screw speed during mixing exhibited strong influence on the conductivity of the nanocomposites above such value. Addition of ~ 2 wt. % GNP into PET matrix caused an increase in the degree of crystallinity, accelerated both cold and melt crystallizations along with good improvements in the thermal stability.
However; Further addition (> 2 wt. %) improves thermal properties slightly. This attributed to higher agglomeration, pulling out and rolling up of GNP sheets during mixing process; which reduced their aspect ratio and then reduced reinforcement efficiency.

Acknowledgement
I would like to thank King Abdulaziz City for Science and Technology (KACST) for their funding throughout this project.

The conflict of interest
The author(s) declare(s) that there is no conflict of interest regarding the publication of this paper